Issue Performance measurement breaks down as knowledge work absorbs AI

Summary

McKinsey argues that measuring performance is getting harder as knowledge work grows and AI becomes part of how that work is done. Conventional performance management looks backward at past results, overlooks work carried out with AI, and often holds back skill development. The report calls for replacing annual ratings with transparent goals that shift over time and for managers who coach in real time. In workplaces that mix people and AI agents, metrics must recognize both human contribution and agent output, so that people remain accountable while agents supply speed, with the two supporting each other. Clear ownership of results, fair rewards for roles that orchestrate agents, and better-trained managers are presented as crucial.